Advances in Science and Technology Vol. 181

Abstract: The engineering historical analysis of Maag, Fellows and Pfauter gear manufacturing processes are important to understand the development of the manufacturing technologies. These unique technologies not only show the different engineering approaches and innovative significances. They show how the pretensions of the manufacturing accuracy, labor productivity and automatizations are developed during the 20th century. The technical historical and functional analysis contribute the deeper understanding of manufacturing mindset methods and it allows the current and future developments into the technical historical background. The technical historical research of gear manufacturing processes is an interdisciplinary research area: the principle is created by the technical history but it is influenced by engineering and technological aspects at the same time. The manufacturing technologies are not only analyzed in technical aspect, but they are also analyzed in social and industrial aspects.

Abstract: This paper presents a comparative study between cycloidal drive components manufactured using Fused Deposition Modeling (FDM) with PLA plastic and CNC laser-cut aluminum. Building upon previous work with polymer-based cycloidal reducers, this research investigates the performance characteristics, mechanical properties, and practical applications of precision-machined aluminum cycloidal discs and input shafts for high-torque gear transmission systems. Experimental testing was conducted on 24:1 reduction ratio cycloidal drives with identical geometries but different materials for the cycloidal disc and input shaft components. Key performance metrics including maximum torque capacity and thermal behavior were measured and analyzed through torque testing and thermal monitoring. Improvements​‍​‌‍​‍‌​‍​‌‍​‍‌ in torque capacity were primarily demonstrated by the aluminum construction (1.7× increase) along with trade-offs in manufacturing cost, weight, and production complexity being identified. The aluminum cycloidal disc achieved 31.5 Nm maximum output torque compared to 18.3 Nm for PLA, with substantially better thermal stability. To verify design decisions and understand structural response, finite element analysis was also conducted. This research serves as a knowledge base filled with real-life scenarios that will aid the engineers to make a confident decision in the choice of manufacturing technology for the next robotic and mechatronic devices ​‍​‌‍​‍‌​‍​‌‍​‍‌projects.

Abstract: Business process discovery aims to reconstruct accurate models of organizational workflowsfrom event logs, yet classical algorithms often struggle with noisy and heterogeneous data,producing overly complex or imprecise models. This paper presents a systematic, quantitative evaluationof clustering-based preprocessing as a means to enhance process discovery. Using the BPIChallenge 2014 incident management log, we applied comprehensive feature engineering to extractstructural, temporal, and behavioral attributes, followed by multiple clustering techniques (K-Means,DBSCAN, HDBSCAN, GMM). Within each cluster, four established discovery algorithms (Alpha Miner,Heuristic Miner, Inductive Miner, ILP Miner) were executed from the PM4Py Python libraryand evaluated using fitness, precision, generalization, and simplicity metrics. Results show that clusteringconsistently improved model simplicity and precision, while fitness remained stable and generalizationtended to decrease. The extent of improvement depended on both the clustering methodand the discovery algorithm: K-Means with higher cluster numbers enhanced fitness and simplicityfor Alpha Miner; HDBSCAN improved precision and simplicity for Heuristic Miner; InductiveMiner benefited from K-Means even with small cluster numbers; whereas ILP Miner models remainedhighly complex regardless of clustering. These findings highlight the trade-off between interpretabilityand generalization, and demonstrate that clustering-based preprocessing can yield moretransparent and actionable process models.

Abstract: Accurately predicting real estate values remains challenging in volatile and rapidly growing real estate markets, especially for expensive homes worldwide. Traditional statistical and economic models frequently miss the temporal dynamics and nonlinear dependencies affecting changes in property values. This study uses a large-scale, multi-market dataset to anticipate real estate values utilizing sophisticated deep learning architectures, including Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), and Dense Neural Networks (DNN). To increase learning efficiency, data pretreatment techniques included time-series sequencing, categorical encoding, and normalization. Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and the Coefficient of Determination (R2) were used to assess the model's performance. The findings show that the LSTM model obtained the lowest average prediction error (MAE) and produced extremely accurate forecasts by effectively incorporating long-term temporal dependencies. According to the study's findings, deep learning gives a solid, scalable foundation for accurate property price predictions. It also has practical ramifications for urban planners, investors, and legislators in dynamic housing markets.

Abstract: Enhancing manufacturing efficiency increasingly relies on integrating Overall Equipment Effectiveness (OEE), Industry 4.0 technologies, and Artificial Intelligence (AI) with Machine Learning (ML). Traditionally, OEE aggregates Availability, Performance, and Quality, but in Industry 4.0 it can evolve into a real-time decision-support tool enriched with predictive analytics. This paper provides a conceptual systematic synthesis of the interplay between OEE and AI-driven methodologies, emphasizing the role of fuzzy logic and hybrid models in managing uncertainty. AI/ML applications—predictive maintenance, quality assurance, and process optimization—reduce downtime, minimize scrap, and increase productivity by improving OEE components through pattern recognition and forecasting. A further research focus is domain shift and transfer learning, which impact the scalability and robustness of industrial AI systems across changing equipment and factories. Transfer learning approaches such as fine-tuning, feature alignment, and adversarial adaptation can reduce retraining effort while maintaining performance. Finally, integrating AI with OEE supports sustainable manufacturing by improving energy efficiency and reducing environmental impacts, contributing to the long-term vision of self-optimizing digital factories capable of responding dynamically to market and environmental changes.

Abstract: The application of virtual factory simulation plays a pivotal role in the design and optimization of modern production systems. By enabling the evaluation and refinement of manufacturing processes prior to physical implementation, process simulation significantly enhances decision-making efficiency. Real-time optimization has attracted considerable attention in the process industry and is widely adopted, as it typically relies on external databases and parameter sets. This study investigates the import, use, and management of such external parameters, providing a comprehensive overview. A detailed case study is developed and solved to demonstrate the proposed approach. The results show that process simulation with parameterized models improves system efficiency and enables real-time optimization capabilities.